While climate change science has advanced considerably in recent years, and evidence of increasing atmospheric carbon is demonstrable, there is yet much unknown about the consequences of anthropogenic carbon emissions. For instance, the effect of a marginal ton of carbon emissions on earth warming is uncertain. The environmental consequences of a marginal increase in the earth's temperature are also uncertain. Similarly, it is not known which mechanisms are the most cost effective to mitigate against climate change or whether adaptation to climate change would be better than mitigation. This problem addresses these uncertainties.
In response to climate change concerns, the government can either invest in mitigation or adaptation or do nothing. Mitigation reduces the flow of carbon into the atmosphere, thereby changing the probability distribution of future climate states. Adaptation reduces the damage costs of any given climate state. Assume the government is considering two projects and can invest $1,000M(illion) in at most one project to control damages from carbon emissions. The two projects are as follows:
Project 1: At time t=0, it is believed that with probability beta=0.75, $1,000M invested in solar and wind power technologies will eliminate power plant emissions entirely, reducing emissions by 60M tons per year. Otherwise, the technology proves less effective and only displaces the most costly fossil fuel power plants so that emissions are only reduced 20M tons per year.
Project 2: The government invests in an adaptation project (such as a sea wall) that costs $1,000M and reduces the damages from climate change. At time t=0, it is believed that with probability delta=0.5 the project reduces damages by 30% and otherwise it reduces damages by 20%.
Assume the interest rate is r=0.10, that costs are born immediately, and that the stream of benefits (and damages) is infinite.
What is the NPV of expected damages if Project 2 is implemented at t=0? What is the expected NPV of Project 2 benefits? Is Project 1 worthwhile?